$4l0,0c)0 85
$4,200
Applications
rules for prediction of future activity.
Recently~,
magnetic drumI
computer was utilized for the technical analysis o( securities., and one stock can be analyzed on this computer in som!ewhat less than ITlany companies. At General Electric's 1ppliance Park plant .in Louisville, Kentucky, a 20-week productio"h, schedule, involving thirty models, 1,000 purchased parts and 350 raw'material items, can be systems have been utilized in business applications. At the present time there are approximately 450 modern electronic computers .being
Actual Installations
used in data process.ing applications for business type problelTls .in the United States; Over fifty of these are of the large scale variety, that is; UNIVACS~ IBM 702's and 70S's. (The'702, incidentally, was the predecessor of the 705). The relTlaining four hundred are of the lTled~
iUlTl scale or'small scale variety and are of the type referred to as magnetic drum compute r s .
16
. Incidentally., the'potentiallTlarket for electronic cOlTlputers has been predicted in a recent study performed for a brokerage firm by a lTlan-agement consulting' company in the following terms: It is stated that there are firm orders for well over 1,000 internally programlTled COITl-puters, and that the value of the backlog is estilTlated to be in excess of
$4000, 000, 000. The total deITland for gene ral purpose digital COITlpU-ters was 'estiITlated in this study as $2.4 billion and the breakdown was approxiITlatel y as .follows: About $1, 250', 000 for the large scale COITl-puters and about $1,150,000 for the ITlediuITl scale cOlTlCOITl-puters.
So you see~ it's about an equal division here. About a billion and a . half~would be·for business applications, $500 lTlillionfor science and
engi~neering, and $350 lTlillion for government. So here we see a ratio of three to one for busines s applications over scienc:e~ and enginee ring.
The early history of the cOlTlputer was almost the' upposite of this. That is, the preponderance of the applications were in the scientific and en-gineering fields.
.I.~.ctual Installations
Benefits to Office Ope rations
I promised at the beginning that we might indicate what the possible benefits to the operations people in today's offices might be from the utilization of electronic data processing systems. The primary ad-vantages seem to be the following:
1. Reduced costs, as we have seen in the insurance application.
This is largely the result of reduced personnel.
2,. Increased speed. That is, the information can be made avail-able to management in reports much more rapid! y than it could have in the old manual and semi-automatic electromechanical data processing systems.
3. Increased accuracy. . The accuracy results from self-checking features, which can be built into the computer, and the ability to perform multiple program checks on its operation, i. e., the inherent accuracy of electronic computers as compared with manual or semi-automatic methods.
4. Improved reporting. It is possible for the first time to obtain new reports consolidated with respect to more sophistacated or complex criteria or including only items which are of an ex-ceptional nature, thet so-called management by exception rou-tines.
5. The consolidation of files. This has been particularly apparent in the insurance data processing applications in which many in-creased accuracy and more efficient operations.
17
Benefits to Office Operations 18
6. . The automatic processing exceptions . In manual systems, excep-. tions we re usually handled without too much difficulty by human
operators who made decisions based upon policy and information
.
.available. However, in punched cards systems it represented quite a bit of additional processing and actually the operations were sufficiently obnoxious in some cases as to be handled or pro-cessed by hand., Electronic data processors can handle such ex-ceptions efficiently and machine decisions are consistent with the policies established, whereas sometimes human interpretations of company policies or procedures can vary.
Future Data Processing Equipment Trends
Pe rhaps it would also be appropriate to have a few remarks about some of the future equipment trends. Some of the maj o~ problems remaining in electronic data processing equipment are the following:
.1)
Sorting is not particularly efficient with magnetic tape compu-ters. It is a time consuming ,operation and indeed some of the applica-tions which I have examined have resulted in a ratio of the time required for sorting to that required for proce s sing to be ten to· one or even great-er. To meet this requirement, several companies are presently develop-ing, an~ actually some of them have available on the market, special purpose sorters which are off-line devices; that is, they don't require the main computer for the sorting but do it with a separate auxiliary device independent of the computer itself.2) Another d~ficiency (in the present magnetic tape systems at least) is the lack of sufficiently rapid random access for answering interrogations. To run through alnagnetic tape file from one end to
Future Data Processing -Equipment Trends 19
To satisfy these requirements Jor interrogations in those applications where 'necessary, a number of companies have been developing random type tube whereby the information is displayed in visually readable char-acters which can be optically r.ecorded on thirty-five millimeter film and
Future Data Processing Equipment Trends 20
later reproduced on hard copy by means of anyone of a number of pro ..
cesses, such as Xerography,.·
Company Program for EDPS
A!few concluding remarks might be appropriate as .to how a company might reasonably start upon a program of investigation of electronic data processing systems; i. e., application possibilities in their partic-uhl.r ,company ~, One way to do this is, to conduct a feasibility study of those areas which seem to be the most attractive. To determine those which ·aremostattracti veaccurately really requires a detailed study;
however, some of the: characteristics can be easily recognized. Those which have a" large volume of repetitive operations quite possibly will prove fruitful for 'electronic data processing.
Such a feasibility study might be performed by one's own employees but frequentl y it proves very useful to bring in an outside consulting firm, which can do an impartial and efficient job of such a feasibility study.
Several of us in the management consulting field have been performing this activity lately. It requires the combination of three talents:
1) the business system analyst; . 2) the .data processing specialist or engineer; and 3) Operations Research talent represented typically by the applied mathematician or scientist. We feel this combination in a team organization represents an :optimum approach to an efficient sys terns study 0
Regardless of what method you utilize to conduct your feasibility study, 'it is 'recommended that it have a high level backing. An electronics
committee should be set up to review its results and to become familiar with the, potentialities of electronics by ,means of an education program.
Company Program for E'DPS
It seems to me that we are now in the third, and perhaps the final stage of the evolution of ele ctronic data proce s sing systems' and appli~ations,'
The first of these three stages was the stage of romance in which there was a tremendous interest in this new and powerful - - if you will; ro-mantic tool - - with its great pos sibilities as an electronic brain, It' was going to solve all problems, create a Utopia, This stage, of course, was characterized by a gross over-statement of the capabilities of this tool; it included a great elation in its potential, only to be followed by a second stage of disillusionment, in which these wonderful promises for electronic s we re not actually realized,
21
However, I think today that we are in the third stage, The pendulum has swung back almost to the middle as, a reaction to the disillusionment. We are finding today a number of very practical, workable installations of electronic data processing systems" Management is beginning, to realize that computers do have a place, perhaps not the exaggerated one that was
\
claimed in the romantic stage, but a very practical place in today's
bus-i
ines s wo rId. This current st~ge might be referred to as one of realism.
I fe'el if we all remember that we do not have the panacea here for all the problems of increasing paper work, and if we thus approach a realistic feasibility study, we should find that electronic data processing systems will prove very beneficial indeed to the mechanization of many of our routine office operations.